Aquatic Invasions (2018) Volume 13, Issue 3: 409–420 DOI: https://doi.org/10.3391/ai.2018.13.3.08 Open Access © 2018 The Author(s). Journal compilation © 2018 REABIC Research Article

Invasion success and population characteristics of the opossum shrimp, Mysis diluviana, in Wyoming, USA

Brett M. Johnson1,*, William M. Pate1, Douglas B. Silver1 and Julia L. Sharp2 1Department of Fish, Wildlife and Conservation Biology, Colorado State University, 1474 Campus Delivery, Fort Collins, CO 80523, USA 2Department of Statistics, Colorado State University, 1877 Campus Delivery, Fort Collins, CO 80523, USA Author e-mails: [email protected] (BMJ), [email protected] (WMP), [email protected] (DBS), [email protected] (JLS) *Corresponding author Received: 9 December 2017 / Accepted: 11 May 2018 / Published online: 9 July 2018 Handling editor: Darragh Woodford

Abstract

Studying the colonization, distribution, demographics, and abundance of invasive species is important for understanding their invasion biology, including the conditions required for establishment. This information can also be used to reduce their risk of spread. Opossum shrimp Mysis diluviana Audzijonyte and Väinölä, 2005, is an invasive species in lakes and reservoirs of the western United States and Canada. Four lakes in western Wyoming, USA, were stocked with this nonnative crustacean in 1971, but no Mysis surveys have been conducted in Wyoming since 1981. We determined presence/absence, demographics, and abundance of Mysis in these and six nearby lakes that could have been invaded using vertical net tows and environmental DNA analysis. Environmental conditions were compared in lakes with and without Mysis, and we evaluated the potential for Mysis to disperse downstream. Mysis (> 500 individuals/m2) persisted in two of the four stocked lakes and nowhere else. Both of the lakes with established populations had daytime light levels on the bottom below the visual feeding threshold for fish, and the hypolimnia were oxygenated. Hierarchical cluster analysis of lake physicochemical conditions grouped these two lakes with four others, all of which were deep (46–185 m), with high oxygen concentrations (> 3 mg/L) on the bottom, and relatively low light intensities (< 0.2 lx) near the bottom. A second cluster of lakes that all lacked Mysis, and appeared to be less suitable, were shallow (< 20 m), had severe hypolimnetic hypoxia, and higher light levels (≥ 590 lx) near the bottom. The interaction of strong light penetration with lake depth, compounded by strong clinograde oxygen profiles, would prevent the formation of a daytime predation refuge from fish in these lakes, reducing the likelihood of Mysis invasion. Given that only half of the purposeful introductions in Wyoming were successful, and that there have been no new invasions in nearly 50 years, future range expansion by the species in the region is unlikely without human facilitation. Key words: Mysidacea, crustacean, dispersal, lakes

Introduction for understanding habitat requirements, dispersal pathways, identifying sites that are vulnerable to Humans have intentionally introduced thousands of secondary spread of the species, and implementing species to the United States and extended the range management actions to minimize this risk (Vander of many more (Pimental et al. 2005). Some of these Zanden and Olden 2008; Stohlgren and Jarnevich species are invasive, meaning capable of spreading 2009). and establishing new self-sustaining populations The opossum shrimp Mysis diluviana, Audzijonyte outside their native range (Kolar and Lodge 2001). and Väinölä, 2005 (formerly M. relicta Lóven, 1862, Some invasive species may also become a threat to herein Mysis), was introduced into many lakes and human health, industry or local biodiversity (Pejchar reservoirs of western North America during the and Mooney 2009). Monitoring the current distribution 1950s–1980s to improve the forage base for salmonid and abundance of harmful invasive species is important sport fishes (Lasenby et al. 1986). Mysis are small

409 B.M. Johnson et al.

(< 25 mm in body length) shrimp-like crustaceans and they are relatively weak swimmers (Ricker 1959), native to deep, oligotrophic lakes in parts of the Mysis dispersal is passive, via river connections to American Great Lakes region and Canada that were downstream lakes, or facilitated by interbasin water glaciated during the Pleistocene (Dadswell 1974; transfers. As with many other aquatic invasive species Audzijonyte and Väinölä 2005). They are coldwater (AIS; Johnson et al. 2001; Wittmann and Ariani 2009), stenotherms with a preferred temperature of about 9 °C Mysis could also be transferred overland to uncon- (Boscarino et al. 2010) and an upper lethal tempera- nected waters by recreational boaters. Despite their ture of 20 °C (Dadswell 1974; Degraeve and Reynolds limited dispersal capabilities, the history of Mysis 1975). Mysis are omnivorous, feeding on organic introductions and subsequent invasions demonstrates matter in the profundal zone during the day and the high vagility of this species. Establishing a migrating into the water column at night to feed on population in Lake Tahoe, California-Nevada, was phytoplankton and zooplankton (Grossnickle 1982; described as “surprisingly easy” (Fredrickson 2017). Beeton and Bowers 1982). A single thermos containing 600–1,000 mysids The initial introduction of Mysis in western North collected from a lake in Minnesota was sufficient to America occurred at Kootenay Lake, British Columbia, establish a massive population in Twin Lakes, Canada in 1949. A dramatic increase in the growth Colorado (Gregg 1976). Twin Lakes then became of kokanee Oncorhynchus nerka Walbaum in Artedi, the source population for subsequent introductions in 1792, prompted well over 100 subsequent introductions > 50 other waters in Colorado and Wyoming (Nesler across western North America (Lasenby et al. 1986). 1986; Martinez and Bergersen 1989). Many of the Most of these introductions proved harmful to fish introductions in Colorado were successful (DBS, populations, as studies showed that their diel vertical unpublished data) and several other water bodies migrations allowed Mysis to avoid predation by were subsequently invaded from upstream waters or planktivorous fish in many lakes, and that Mysis facilitated by interbasin water transfers (Nesler consumed the zooplankton preferred by these fishes 1986; DBS, unpublished data). Given the high (Nesler and Bergersen 1991). Thus, instead of beco- vagility of Mysis and their potential to harm invaded ming a new food resource, introduced Mysis often systems, determining invasion success, monitoring became strong competitors and the growth of sport population sizes and determining the extent of their fish such as kokanee and rainbow trout O. mykiss range is important for assessing risk and preventing Walbaum, 1792, declined (Lasenby et al. 1986; Nesler future Mysis invasions. and Bergersen 1991). Mysis introductions also stimu- Although most Mysis introductions in this region lated reproduction of nonnative lake trout Salvelinus occurred about 50 years ago, there have been very namaycush Walbaum in Artedi, 1792, which preyed few studies to determine the outcome of many of on other sport fish and native species leading to fish these introductions, or surveys to investigate if there population declines (Martinez et al. 2009; Schoen et have been subsequent invasions. In Wyoming, there al. 2015). Introduced Mysis have precipitated even are a multitude of lakes and reservoirs with suitable broader food web changes. Selective predation by Mysis habitat, but very few surveys of the species in Mysis on large zooplankton favors smaller-bodied the state. Mysis were introduced to four lakes in species (Martinez and Bergersen 1991) which are western Wyoming in 1971. Half Moon Lake, Middle less efficient grazers of phytoplankton (Brooks and Piney Lake, and Willow Lake near Pinedale, Wyoming Dodson 1965). The diel vertical migrations of Mysis were stocked with Mysis obtained from Twin Lakes, have affected contaminant and nutrient transport and Colorado (Finnell 1977). Approximately 14,000 indi- recycling rates (Van Duyn-Henderson and Lasenby viduals were stocked in Half Moon Lake and Willow 1996; Devlin et al. 2016). Mysis-induced changes to Lake in June, 1971 and an additional 50,000 were the biomass and composition of the zooplankton stocked in each lake in October (Grabowski and community and disturbance of nutrient dynamics Ahern 1982). Approximately 20,000 mysids were have altered primary production of recipient systems stocked in Middle Piney Lake that year as well. (Devlin et al. 2016). Thus, knowledge of Mysis distri- Finnell (1972) stated that Brooks Lake, northwest of bution and abundance can help managers determine Dubois, Wyoming, was also stocked with Mysis but appropriate stocking and harvest policies for both no details about numbers or dates stocked were planktivorous and piscivorous sport fish, evaluate lentic provided. Surveys conducted by Wyoming Game fish conservation strategies and interpret fish commu- and Fish Department (WGFD) in 1976 and 1977 and nity and water quality dynamics (Ellis et al. 2011). by the University of Wyoming in 1981 confirmed Once Mysis are established in a region, they can that Mysis had established self-sustaining populations spread to nearby waters (Nesler 1986; Spencer et al. in Half Moon and Willow lakes (Grabowski and 1991). Because they are an obligate lacustrine species Ahern 1982).

410 Mysis diluviana invasion success

Table 1. Characteristics of the study lakes, near Pinedale, Wyoming, USA. All of the lakes are within the Green River Basin except Brooks Lake, which is in the Wind River Basin. “Stocked?” refers to whether Mysis were introduced by management agencies. Lake Latitude N Longitude E Elevation (m) Surface area (ha) Maximum depth (m) Stocked? Boulder Lake 42.847 −109.664 2,222 700 79 N Brooks Lake 43.757 −110.004 2,764 87 18 Y Fremont Lake 42.946 −109.806 2,261 2,061 185 N Half Moon Lake 42.925 −109.730 2,316 381 85 Y Little Half Moon Lake 42.905 −109.710 2,315 24 17 N Meadow Lake 42.888 −109.689 2,434 50 20 N Middle Piney Lake 42.598 −110.574 2,699 67 46 Y New Fork Lake 43.097 −109.947 2,383 497 62 N Soda Lake 42.961 −109.844 2,302 129 14 N Willow Lake 43.002 −109.869 2,346 726 85 Y

We are unaware of any other surveys of Mysis popu- lations in Wyoming until the present study. The objectives of this study were to determine the presence/absence, demographics, and abundance of Mysis in the four lakes where they were originally introduced, and evaluate invasion success. We also sampled six neighboring lakes that could have been colonized by the stocked populations. We compared environmental conditions in lakes with and without Mysis to better understand habitat requirements for this invasive species outside its native range. Finally, we evaluated the potential for Mysis to disperse to downstream waters.

Methods The survey was conducted on 10 glacial lakes in the Bridger-Teton and Shoshone National forests, in northwestern Wyoming, USA (Figure 1). During July 17–20, 2017 we sampled the four lakes reported to have been the original Mysis introduction sites (Brooks, Half Moon, Middle Piney, and Willow lakes). We then sampled six neighboring lakes during August 28–31, 2017 (Boulder, Fremont, Little Half Moon, Meadow, New Fork, and Soda lakes) which were all within 10 km of a lake inhabited by Mysis because other crustaceans can disperse readily over this distance by a variety of mechanisms (Havel and Shurin 2004). These lakes were also > 10 m deep and therefore likely to stratify, providing cold hypo- limnetic temperatures required by Mysis. Each of the lakes was also directly accessible by public roads that made overland transfer of Mysis via trailered boats more likely. The lakes were relatively small (mean: Figure 1. Locations of 10 Wyoming waters surveyed for Mysis 477 ha, range: 24–2,061 ha), deep (mean: 61 m, range: diluviana: 1) Boulder, 2) Brooks, 3) Fremont, 4) Half Moon, 14–185 m), high elevation (≥ 2,222 m AMSL) waters 5) Little Half Moon, 6) Meadow, 7) Middle Piney, 8) New Fork, (Table 1). Only Soda Lake did not have a surface 9) Soda, and 10) Willow lakes. Location of Fontenelle Reservoir (11), which was not surveyed but is connected to study lakes, is also shown. outflow. All of the lakes are managed as coldwater fisheries with a variety of salmonid species including brown trout Salmo trutta Linnaeus, 1758, cutthroat

411 B.M. Johnson et al. trout O. clarkii Richardson, 1836, lake trout, and Middle Piney, and Willow lakes) or if Mysis were rainbow trout. Native fishes include highly unusual detected in our vertical net tows or eDNA sampling. lacustrine populations of two imperiled riverine The remaining lakes were classified as “exposure species in some of the lakes, flannelmouth sucker unknown” because we could not know if Mysis had Catostomus latipinnis Baird and Girard, 1853, and ever been present in the past but did not persist. roundtail chub Gila robusta Baird and Girard, 1853 We measured environmental conditions at one to (Laske et al. 2011), as well as mountain sucker Cato- five stations on each lake, depending on the parameter stomus platyrhynchus Cope, 1874, and speckled dace measured and lake area. A temperature-dissolved Rhinichthys osculus Girard, 1856. oxygen profile was measured at the deepest station We used a stratified sampling design to establish in each lake. Other water quality parameters were Mysis sampling stations on each lake (Martinez et al. measured at the surface at three stations (area < 400 ha) 2010). The number of stations chosen depended on or five stations (area ≥ 400 ha) per lake. Surface con- lake area (area < 40 ha: 3 stations, 40 ≤ area < 400 ha: ductivity, pH, salinity, and TDS were measured with 5 stations, area ≥ 400 ha: 10 stations). Within each an Oakton PCSTestr35 multimeter. Turbidity was lake, stations were distributed across up to four depth measured at the surface with a Hach 2100Q turbidi- contour strata: 10–20 m, 20–40 m, 40–60 m, and meter. Water transparency was measured during > 60 m, as lake depth (Zmax) allowed. Three sampling daylight with a standard 20-cm Secchi disk on the stations were established on Little Half Moon Lake, shaded side of the boat without sunglasses. Tempe- five on Brooks, Half Moon, Middle Piney, New Fork rature and dissolved oxygen profiles were measured and Soda lakes, and 10 stations on the larger Boulder, with an YSI ProODO meter. We evaluated suitability Fremont and Willow lakes. A water sample from the of thermal and oxygen conditions assuming that hypolimnion of each lake was collected with a van Mysis would avoid temperatures ≥ 14 °C (Ricker 1959; Dorn bottle for Mysis environmental DNA (eDNA) Martinez and Bergersen 1991) if dissolved oxygen analysis. We have used this eDNA method in con- allowed (Dadswell 1974; Degraeve and Reynolds junction with vertical net tows in at least 20 lakes 1975), and that the avoidance threshold for hypoxia with no discrepancies in presence/absence between in Mysis was 3 mg/L (Sandeman and Lasenby 1980; the two sampling methods (BMJ, unpublished data). Sherman et al. 1987). We used temperature and dis- We could not sample Meadow or Soda lakes with solved oxygen profile data and this oxygen threshold the Mysis net due to time constraints, but these two to determine the maximum habitable depth (Z3), and lakes were shallow and had the most severe hypoxia the corresponding temperature at that depth (T3), in which suggested that it was unlikely Mysis could each lake, because low dissolved oxygen in the persist there. We did collect eDNA samples at these hypolimnion has been shown to force a closely related two lakes to determine presence/absence. species (M. relicta) into unfavorable water tempera- Mysis sampling commenced at least 60 minutes tures (Horppila et al. 2003). after sundown and consisted of vertical tows with a Our assessment of habitat suitability also depended 1-m diameter plankton net with 500-µm Nitex mesh on the presence of a daytime predation refuge. We (Silver et al. 2016). Boat lights were turned off during used water transparency and surface illuminance to vertical net tows to avoid affecting Mysis distribution. compute light availability at the bottom of each lake The net was lowered until the cup was within 1 m of to determine if Mysis have a low-light refuge from the bottom, as guided by a depth sounder. The net visual predators (fish), in the manner of Hansen and was allowed to rest for 60 s and then retrieved at a Beauchamp (2015). A light extinction coefficient, k, constant rate of 0.4 m/s with an electric winch. We was computed from water transparency (Idso and collected one sample at each station. The catch from Gilbert 1974): each haul was preserved in 70% ethanol. The water k= 1.7/Z for the eDNA samples was collected with a Van Dorn SD sampler from within 1 m of the bottom and processed where ZSD is Secchi depth (m). The mean June– using the protocol provided by Carim and Wilcox August surface illuminance, I0 (lx), during midday (2014). Prior to collecting the lake water eDNA sample was computed from an algorithm that used latitude, a distilled water control was processed to insure longitude, date, and time of day (Janiczek and adequate decontamination of the sampling equipment. DeYoung 1987). We estimated illuminance at depth The eDNA samples were analyzed by the U.S.G.S. from surface illuminance and the extinction coefficient Molecular Ecology Laboratory in Fort Collins, using the Beer-Lambert equation (Horne and Goldman Colorado, using methods described in Carim et al. 1994): (2016). Lakes were classified as “exposed” if they -kZ were known to have been stocked (Brooks, Half Moon, IZ=I0•e

412 Mysis diluviana invasion success

where IZ is illuminance (lx) at depth Z (m). We Results computed light intensity at the maximum habitable depth for Mysis (dissolved oxygen ≥ 3 mg/L), I3, We captured Mysis in only three lakes: Half Moon, because low dissolved oxygen in the hypolimnion Little Half Moon, and Willow lakes. The eDNA samples could have forced Mysis into shallower water where yielded positive detections in these three waters and they would be visible to predators. A predation no others. Net and eDNA results suggest that Mysis refuge was assumed to exist if I3 < 0.001 lx, the have established only at Half Moon and Willow minimum light intensity for visual feeding in lakes. We captured a single juvenile (9.3 mm) mysid salmonids (Ali 1959). in our sampling at Little Half Moon Lake for an In the laboratory, the total catch from each net estimated density of 0.42 individuals/m2. In contrast, haul was enumerated directly, or the catch was we captured 3,080 mysids at Half Moon Lake, for an subsampled with a Folsom plankton splitter (Sell areal density of 784.7 individuals/m2 (SD = 533.5) and Evans 1982). Subsamples of approximately 150 and 4,612 mysids at Willow Lake, for an areal density mysids from each lake and station were examined of 587.5 individuals/m2 (SD = 522.4). Mysis density under a stereomicroscope at 7x magnification. Each was not statistically different at Half Moon and individual was classified as 1) juvenile (< 10 mm), Willow lakes (t(13) = 0.69, P = 0.49), nor were these 2) male (extended pleopods), 3) female (brood pouch densities different from those reported by Grabowski exposed), or 4) adult of undetermined sex (≥ 10 mm and Ahern (1982) for Half Moon and Willow lakes and lacking identifiable sexual characteristics). Each (t(6) = −0.28, P = 0.78; and t(11) = −0.06, P = 0.95, mysid was measured (nearest 0.1 mm) along a dorsal respectively). Juveniles comprised 41.97% of the line from the tip of the rostrum to the tip of the total population in Half Moon Lake, and 64.23% in telson using a calibrated micrometer. Lengths in the Willow Lake (Figure 2). The modal lengths of juveniles measured subsamples were weighted by total catch and adults were larger in Willow Lake compared to at each station to create an overall length-frequency Half Moon Lake, suggesting that growth may be distribution for the population. Total counts of the catch faster in Willow Lake. The maximum size of adults in each sample were normalized to individuals/m2 was similar in the two lakes (20.63 mm in Half Moon based on the cross-sectional area of the net opening Lake, 20.81 mm in Willow Lake). In both lakes, all (0.785 m2). of the males were immature and over 99% of females Areal densities of Mysis were compared with were immature, as expected for the time of year. unpaired t-tests (equal variances) in the two lakes All of the lakes were thermally stratified (Figure 3). with established populations of Mysis. We also Surface temperatures were lowest at Brooks (15.4 °C) compared our estimated densities in these two lakes and Middle Piney lakes (14.7 °C) in July; these lakes with those in a previous study (Grabowski and were 300–500 m higher in elevation than the others. Ahern 1982) with two-sample t-tests as an indication Surface temperatures of the other two lakes sampled of population stability. Physicochemical measures in July were 18.9 °C (Half Moon Lake) and 20.7 °C taken at multiple stations in each lake were averaged (Willow Lake). Surface temperatures were more so that each lake had one observation. Correlations similar across the six lakes sampled in August among physicochemical characteristics (area, con- (17.3 °C–19.8 °C). The top of the thermocline was ductivity, dissolved oxygen on the bottom, I3, pH, shallower in the lakes sampled in July (~ 4 m) than salinity, T3, total dissolved solids, turbidity, Z3, Zmax, those sampled in August (5–9 m). The temperature and ZSD) across lakes were evaluated with the of the hypolimnion was 4–6 °C in all of the lakes Pearson product-moment correlation coefficient, r, except one; Little Half Moon Lake, which is relatively prior to selecting variables for the cluster analysis. A shallow and receives inflow from the surface of Half Bonferroni correction was used for multiple compa- Moon Lake, was relatively weakly stratified and had risons; a significance level of 0.0009 was used to a bottom temperature of 15.2 °C. judge significant correlations (α = 0.05). A subset of Dissolved oxygen concentrations at the surface uncorrelated variables (conductivity, DO, I3, pH, were high (≥ 7.5 mg/L) in all of the lakes (Figure 3). turbidity, Z3, and ZSD) was used in a hierarchical Dissolved oxygen concentration was above the Mysis cluster analysis to examine physicochemical avoidance threshold (3 mg/L) throughout the water characteristics among the 10 lakes, and associations column in all of the lakes except in the four shallowest of environmental conditions with presence/absence lakes: Brooks, Little Half Moon, Meadow, and Soda of Mysis. Ward’s minimum variance method was lakes. In these four lakes the dissolved oxygen con- used to form the clusters. Two clusters were selected. centration decreased rapidly with depth below the JMP Pro version 13 was used for correlation and thermocline, and the dissolved oxygen concentration cluster analyses. at the bottom was < 0.6 mg/L. Based on the dissolved

413 B.M. Johnson et al.

Figure 2. Size-frequency histograms of Mysis diluviana captured at A) Half Moon Lake on July 17, 2017, and B) Willow Lake on July 18, 2017, near Pinedale, Wyoming, USA.

Figure 3. Temperature and dissolved oxygen profiles measured at a mid-lake station on four Wyoming lakes in July, 2017 (Upper panels), and six Wyoming lakes in August, 2017 (lower panels). Mysis were established in Half Moon and Willow lakes (black lines).

414 Mysis diluviana invasion success

Table 2. Mean and standard deviation (SD) of physicochemical characteristics measured at the surface of 10 Wyoming lakes sampled in July and August, 2017. Conductivity Date pH Salinity (mg/L) TDS (ppm) Lake Stations (µS/cm) sampled Mean SD Mean SD Mean SD Mean SD Boulder Lake 8/29/2017 5 17.90 0.07 7.80 0.08 15.00 0.20 12.70 0.07 Brooks Lake 7/20/2017 3 63.00 0.10 8.83 0.04 33.90 0.30 44.70 0.00 Fremont Lake 8/30/2017 5 18.76 0.46 7.45 0.13 15.70 0.23 13.44 0.38 Half Moon Lake 7/17/2017 3 18.00 0.10 8.27 0.12 15.13 0.12 12.80 0.10 Little Half Moon Lake 8/31/2017 3 17.80 0.10 7.49 0.03 15.37 0.06 12.67 0.06 Meadow Lake 8/31/2017 3 83.37 0.50 8.08 0.03 45.90 0.10 59.27 0.15 Middle Piney Lake 7/19/2017 3 309.33 0.58 8.71 0.01 154.67 1.15 218.33 0.58 New Fork Lake 8/28/2017 3 36.83 1.17 7.97 0.02 23.93 1.07 26.03 1.15 Soda Lake 8/29/2017 3 5.27 0.02 8.75 0.08 3.10 0.02 3.74 0.02 Willow Lake 7/18/2017 5 30.12 0.29 7.80 0.04 21.30 0.14 21.40 0.25

Table 3. Water transparency and light characteristics of the study lakes: ZSD is Secchi depth, k Z3 is the maximum depth habitable for Mysis (dissolved oxygen ≥ 3 mg/L), k is a light extinction coefficient (unitless), and I3 is light intensity at the maximum habitable depth. A low- light predation refuge existed in lakes where I3 < 0.001 lx, the minimum threshold for fish visual feeding.

Lake Turbidity (NTU) ZSD (m) Z3 k I3 (lx) Predation refuge? Boulder Lake 0.76 8.83 79 0.192 2.63 × 10-2 N Brooks Lake 1.80 3.93 12 0.432 5.89 × 102 N Fremont Lake 0.77 8.95 185 0.190 5.78 × 10-11 Y Half Moon Lake 1.16 5.30 85 0.321 1.52 × 10-7 Y Little Half Moon Lake 0.88 4.90 9 0.347 4.64 × 103 N Meadow Lake 1.48 3.70 5 0.459 1.06 × 104 N Middle Piney Lake 0.92 5.15 46 0.330 2.68 × 10-2 N New Fork Lake 0.71 8.05 62 0.211 2.17 × 10-1 N Soda Lake 2.19 4.41 7 0.386 7.09 × 103 N Willow Lake 0.84 5.24 85 0.324 1.11 × 10-7 Y

oxygen threshold, the maximum depths that Mysis greatest depth habitable for Mysis (I3) (Table 3). would inhabit (Z3) were 7 m in Meadow Lake, 8 m in Only three lakes (Half Moon, Willow and Fremont) Soda Lake, 10 m in Little Half Moon Lake, and 13 m had daytime light levels at habitable depths that were in Brooks Lake. Temperatures at these lake depths (T3) below the minimum illuminance required for visual were: 12 °C, 14.5 °C, 16 °C and 6 °C, respectively. feeding by fishes, and thus provided a low-light None of these temperatures exceeds the upper lethal predation refuge for Mysis. Both of the lakes with temperature for Mysis but all were above their preferred established Mysis populations (Half Moon, Willow) temperature and two exceeded the avoidance threshold had a daytime predation refuge. for Mysis. Conductivity, salinity and TDS were significantly Conductivity, salinity and TDS were generally low correlated (r > 0.99, P < 0.0001) among lakes (Table 4). (mean = 60 µS/cm, 34 mg/L, and 46 mg/L, respec- Surface area, Z3, and Zmax were significantly correlated tively) and similar across lakes except for Middle (r > 0.95, P < 0.0001). T3 and I3 were significantly Piney Lake where values of each were much higher correlated (r = 0.90, P = 0.0004). The physicochemical (Table 2). With the exception of Little Half Moon variables used in the hierarchical cluster analysis Lake, ZSD was lower and turbidity was higher in the were conductivity, DO, I3, pH, turbidity, Z3, and ZSD. shallowest lakes. ZSD averaged 5.9 m across the lakes Two clusters were selected from the hierarchical and was lowest in Brooks Lake (3.9 ± 0.6 m) and cluster analysis to characterize lakes (Figure 4). The highest at Fremont Lake (8.9 ± 0.5 m) (Table 3). first cluster consisted of six lakes (i.e., Boulder, New Turbidity was < 1.00 NTU in 6 of the 10 lakes, Fork, Half Moon, Willow, Fremont, and Middle turbidity averaged 1.15 NTU and was highest in Piney) and was characterized by higher conductivity, Soda Lake (2.19 ± 0.54 NTU), and lowest in New higher dissolved oxygen, lower I3, lower pH, higher Fork Lake (0.71 ± 0.60 NTU). Differences in water ZSD, lower turbidity, and higher Z3 than the four transparency and lake depth resulted in 13 orders of lakes in cluster two (Brooks, Soda, Little Half Moon, magnitude differences in estimated light levels at the and Meadow) (Table 5).

415 B.M. Johnson et al.

Table 4. Pearson product-moment correlation coefficients among lake characteristics measured at 10 Wyoming lakes sampled for Mysis diluviana. Area = surface area, Cond = conductivity, DO = dissolved oxygen on the bottom, I3 = light intensity at 3 mg/L of oxygen, Sal = salinity, T3 = temperature at 3 mg/L of oxygen, TDS = total dissolved solids, Turb = turbidity, and Z3 = depth at 3 mg/L of oxygen, Zmax = maximum depth, and ZSD = Secchi depth. Significant correlations (Bonferroni adjusted significance level of 0.0009) denoted in bold text.

Area Cond DO I3 pH Sal T3 TDS Turb Z3 Zmax ZSD Area 1.000 Cond −0.308 1.000 DO 0.672 0.076 1.000

I3 −0.437 −0.107 −0.738 1.000 pH −0.568 0.435 −0.342 0.110 1.000 Sal −0.292 0.998 0.100 −0.116 0.408 1.000

T3 −0.481 −0.238 −0.768 0.902 0.013 −0.245 1.000 TDS −0.308 1.000 0.075 −0.106 0.435 0.998 −0.238 1.000 Turb −0.464 −0.101 −0.685 0.570 0.734 −0.132 0.524 −0.101 1.000

Z3 0.953 −0.168 0.843 −0.606 −0.518 −0.150 −0.653 −0.168 −0.590 1.000 Zmax 0.965 −0.187 0.816 −0.553 −0.538 −0.168 −0.611 −0.187 −0.570 0.997 1.000 ZSD 0.754 −0.230 0.763 −0.560 −0.565 −0.202 −0.528 −0.231 −0.676 0.762 0.747 1.000

Table 5. Hierarchical cluster analysis results (mean, standard deviation) using conductivity (µS/cm), dissolved oxygen on the bottom (DO, mg/L), light intensity at ≥3 mg/L of dissolved oxygen (I3, lx), pH, turbidity (NTU), depth at 3 mg/L of oxygen (Z3), and Secchi depth (ZSD, m). Number of Cluster Conductivity DO I pH Turbidity Z Z observations 3 3 SD 1 6 71.82 (116.60) 6.22 (1.68) 0.05 (0.09) 8.02 (0.43) 0.86 (0.16) 90.33 (48.82) 6.92 (1.88) 2 4 42.38 (36.91) 0.24 (0.28) 5730.90 (4209.89) 8.30 (0.63) 1.59 (0.55) 8.25 (2.99) 4.24 (0.53)

Figure 4. Hierarchical cluster analysis dendrogram using seven physicochemical characteristics of the 10 lakes sampled for Mysis diluviana highlighting lakes grouped into two primary clusters (red and green).

Discussion 3 mg/L, and a daytime predation refuge was available, in both lakes during both periods. Surface Mysis persisted in two of the four lakes where they conductivities in 1981 were slightly lower (16 µS/cm), were originally introduced. The Mysis populations in and Secchi depths slightly greater (8 m) than in Half Moon and Willow lakes have been relatively 2017, but natural seasonal and interannual variation stable since they were introduced, perhaps because in these parameters may explain some of the diffe- physicochemical conditions have not changed greatly rence. There were few historical data on limnological over the period. Grabowski and Ahern (1982) conditions on the other lakes but the available data sampled Mysis in Half Moon and Willow lakes with suggest that lake conditions in the area have been comparable methods in August 1981 yielding density stable for many years. Brooks Lake has experienced estimates (895 individuals/m2 and 608 individuals/m2, hypolimnetic hypoxia since at least 2001 (WDEQ respectively) that were not significantly different 2015). Dissolved oxygen profiles and conductivity from ours. Dissolved oxygen at 60 m was well above in Fremont and New Forks lakes in August 1984

416 Mysis diluviana invasion success

(Peterson et al. 1987) were similar to our measure- and New Fork lakes. The cluster analysis also showed ments, but surface temperatures may have increased that Middle Piney Lake was most dissimilar to the in Fremont, Half Moon and New Fork lakes since other lakes in cluster 1. This suggests that conditions the 1970s (Leopold 1980). may be less suitable for Mysis here, and may also The densities of Mysis in Half Moon and Willow explain why the initial introduction failed at Middle lakes were relatively high compared to some other Piney Lake. We believe there are two plausible introduced populations of the species in western North mechanisms. Mysids are highly sensitive to rapid America. Mysis density averaged 288 individuals/m2 changes in conductivity (Fürst 1965; Gosho 1975). (range: 1–495 individuals/m2) at 15 Colorado reservoirs The conductivity at Middle Piney Lake was at least sampled during 1991–2017 (Martinez et al. 2010; BMJ, three times higher than at any of the other lakes in unpublished data), and averaged 51 individuals/m2 the study, or at the source lake (Twin Lakes, CO; (range: 1–129 individuals/m2) over 30 years at Flathead Britton and Wentz 1980). Thus, stocked mysids may Lake, Montana (Devlin et al. 2016). Seasonal density have died from osmotic shock upon release. A second estimates averaged 352 individuals/m2 (range: 56– explanation is the lack of daytime predation refuge 2,471 individuals/m2) at Lake Pend Oreille, Idaho in this, the shallowest lake in cluster 1. Lake trout, a (Caldwell and Wilhelm 2012). Morgan and Threlkeld predominant predator on Mysis in its introduced range (1982) reported 319 individuals/m2 in Lake Tahoe, (Lasenby et al. 1986), are also more abundant in California. Our densities are also higher than average Middle Piney Lake than in other area lakes (P.A. densities in the Laurentian Great Lakes (≤ 299 Cavali, Wyoming Game and Fish Department, individuals/m2), where Mysis are native (Jude et al. personal communication). Thus, it is possible that 2018). It is difficult to say why Mysis densities are fish predation may have prevented establishment of so high in the Wyoming lakes without more detailed Mysis at Middle Piney Lake. information about phytoplankton and zooplankton The four lakes in cluster 2 (Brooks, Little Half production, and the intensity of fish predation. Moon, Meadow, and Soda) did not have established In addition to updating the distribution of Mysis Mysis populations, although one lake (Brooks Lake) diluviana in Wyoming, our investigation provides was stocked and Mysis were immigrating to another some insights into the invasion biology of this (Little Half Moon Lake). All of these lakes appeared species, including the conditions needed for estab- to be less suitable for Mysis than the lakes in cluster 1: lishment, persistence and secondary invasions. The they were shallow, exhibited low dissolved oxygen likelihood of establishment and persistence were in the hypolimnion, and had a higher pH than lakes higher when Mysis entered systems with physico- in cluster 1. These conditions were negatively asso- chemical conditions similar to those typical of the ciated with the presence of Mysis in 327 lakes in its species’ native range: deep, oligotrophic lakes with a native range (Dadswell 1974). Further, because they well-oxygenated hypolimnion (Dadswell 1974). were shallow and moderately clear lakes, Mysis would Further, we found that the only lakes with estab- be forced to occupy depths that were well above the lished populations of Mysis also had very low light minimum light intensity for fish feeding, increasing intensities near the bottom. Despite relatively high predation risk. Low hypolimnetic dissolved oxygen water transparency, the lakes with Mysis (Half Moon caused obligatory habitat shifts in closely related Lake, Willow Lake) were deep enough (85 m) that M. relicta, resulting in population collapse due to Mysis could retreat to water below the light threshold intense predation by fish (Horppila et al. 2003) for visual-feeding by fish during daytime. Only one which may explain the failure of the introduction at other lake in the study (Fremont Lake) had a daytime Brooks Lake: an interaction between unsuitable predation refuge. oxygen conditions, light penetration and fish preda- The two lakes with established Mysis populations tion. These conditions could also prevent Mysis from clustered together, and these lakes were nested within establishing in Meadow and Soda lakes, if the lakes a cluster with four other lakes, suggesting that were ever exposed to Mysis. Although we captured physicochemical conditions might also be suitable one mysid in Little Half Moon Lake, we believe that for Mysis in Boulder, Fremont, Middle Piney, and Mysis have not established a population there because New Fork lakes. All of these lakes were deeper and surface temperature exceeded the upper lethal had higher dissolved oxygen concentrations in the temperature and depths with suitable temperatures hypolimnion, lower pH, and higher water clarity than were hypoxic. We believe the single individual we the lakes in cluster 2. Light intensity at the bottom was captured had emigrated from Half Moon Lake, just at least two orders of magnitude lower in the lakes in 400 m upstream. cluster 1 than the lakes in cluster 2, but still above The fact that Mysis have not invaded the other the fish feeding threshold at Boulder, Middle Piney, lakes in the study area should be some consolation to

417 B.M. Johnson et al. managers in the region charged with conserving survive the turbulence and fish predation in these imperiled native species in the face of growing rivers to traverse these distances (Gregg and Bergersen numbers of nonnative species. However, just as the 1980). In Colorado, which has dozens of introduced invasion of virile crayfish Orconectes virilis Hagen, Mysis populations, the only cases of successful long 1870, facilitated increased piscivory of native fishes distance (> 20 km) dispersal among interconnected by nonnative smallmouth bass Micropterus dolomieu lakes occurs via pipelines, not surface rivers. More- Lacepède, 1802, elsewhere in the region (Martinez over, mysids dispersing from the Wyoming study 2012), the presence of Mysis in some of the lakes lakes would encounter rapidly increasing and an order may facilitate the invasion of other species, such as of magnitude higher conductivity and salinity as they burbot Lota lota Linnaeus, 1758. Burbot are highly moved downstream in the Green River (Godwin et piscivorous as adults but Mysis are an important food al. 2015). Such changes in water chemistry have for their young (Scott and Crossman 1973) as they are been shown to inhibit translocation success of for young lake trout (Schoen et al. 2015). Burbot are M. diluviana (Gosho 1975) and M. relicta (Fürst 1965). expanding their range in Wyoming and threatening We found no published accounts, agency reports, native fishes, after having been illegally transplanted or database entries documenting sightings of Mysis from across the Continental Divide into Big Sandy in any other waters of Wyoming (B. Bear, AIS and Fontenelle reservoirs during the 1990s (Gardunio Coordinator, WGFD, personal communication; L. et al. 2011). Although burbot have not been detected Tronstad, Lead Invertebrate Zoologist, University of in the Upper Green River basin, there are no barriers Wyoming, personal communication). The absence of preventing them from moving upstream into Little Half evidence for established populations outside of the Moon and Half Moon lakes (P. Cavalli, Wyoming stocked lakes during the nearly 50 years that the Game and Fish Department, personal communication). species has been present in Wyoming strongly If burbot become established in the Half Moon Lake suggests that Mysis have not expanded their range in system they would further threaten the unique popu- Wyoming. Thus, it appears that the prospects for lation of roundtail chub there. Thus, the presence of secondary invasions by Mysis in Wyoming are introduced Mysis could set the stage for an invasional limited without human assistance. Maintaining the meltdown that would amplify impacts on native state’s strong AIS regulations and penalties for species (Simberloff and Von Holle 1999). Because it illegal stocking (WGFD 2017) should help to reduce would not be feasible to eliminate Mysis (Lasenby et the risk of human-facilitated invasions of Mysis in al. 1986; Martinez and Bergersen 1989) conserving Wyoming. Future studies could formally evaluate the native fishes of these natural lakes will depend the risk of Mysis establishing in other waters in on preventing the spread of Mysis and the arrival of Wyoming, and elsewhere in the region. Based on our new invasive organisms. findings, there are several environmental characteristics Interconnected lakes and reservoirs can act as that may increase this risk. Lakes with short stepping stones for invasive aquatic species (Havel connections to established populations seem to be et al. 2005) and another invasive mysid (Hemimysis more invasible to downstream dispersants. The anomala) is currently spreading across Europe and presence of a daytime predation refuge from fishes, the Great Lakes region of North America via inland and cold (< 14 °C) hypolimnetic temperatures coupled waterways (Kestrup and Ricciardi 2008; Minchin with adequate (> 3 mg/L) dissolved oxygen concentra- and Boelens 2010). We believe that Mysis from Half tions also seem to favor colonization and persistence Moon Lake are regularly being flushed into Little of Mysis in Wyoming. A more comprehensive Half Moon Lake and rivers downstream. Mysis may comparison of limnological conditions in lakes where be emigrating to these rivers via the outlet at Willow Mysis invasions have succeeded and failed across Lake also. The outlets from these lakes flow into the their introduced range would provide more insights headwaters of the Green River which has two large, into habitat requirements and a better understanding coldwater reservoirs that could support Mysis appro- of future invasion risk. ximately 110 km (Fontenelle Reservoir), and 220 km (Flaming Gorge Reservoir) downstream. Thus, waters downstream of the Mysis lakes in Wyoming could, Acknowledgements in principle, experience secondary invasion. There We thank Hilda Sexauer for granting permission to sample the lakes, are some reasons to expect this risk to be low. Unlike and Travis Neebling, Craig Amadio, Joe Deromedi, and Pete Cavalli Hemimysis anomala, M. diluviana is an obligate for providing Wyoming Game and Fish Department reports. Dr. lacustrine species and it is improbable that Mysis Sarah Oyler-McCance provided eDNA analyses. This work was supported by a private gift to Colorado State University, account diluviana could survive such extensive transit in a number 6-464540. riverine environment. It is unlikely that Mysis could

418 Mysis diluviana invasion success

References Gardunio EI, Myrick CA, Ridenour RA, Keith RM, Amadio CJ (2011) Invasion of illegally introduced Burbot in the upper Colorado River Basin, USA. Journal of Applied Ichthyology 27: Ali MA (1959) The ocular structure, retinomotor and photo- 36–42, https://doi.org/10.1111/j.1439-0426.2011.01841.x behavioral responses of juvenile Pacific salmon. Canadian Godwin BL, Albeke SE, Bergman HL, Walters A, Ben-David M Journal of Zoology 37: 965–996, https://doi.org/10.1139/z59-092 (2015) Density of river otters (Lontra canadensis) in relation to Audzijonyte A, Väinölä R (2005) Diversity and distributions of energy development in the Green River Basin, Wyoming. circumpolar fresh- and brackish-water Mysis (Crustacea: Science of the Total Environment 532: 780–790, https://doi.org/ Mysida): descriptions of M. relicta Lovén, 1862, M. salemaai n. 10.1016/j.scitotenv.2015.06.058 sp., M. segerstralei n. sp. and M. diluviana n. sp., based on Gosho ME (1975) The introduction of Mysis relicta into freshwater molecular and morphological characters. Hydrobiologia 544: lakes. University of Washington, Seattle. Fisheries Research 89–141, https://doi.org/10.1007/s10750-004-8337-7 Institute Circular No. 75-2, 70 pp Beeton AM, Bowers JA (1982) Vertical migration of Mysis relicta Grabowski JJ, Ahern J (1982) Evaluation of limnological parameters Loven. Hydrobiologia 93: 53–61, https://doi.org/10.1007/BF0000 as related to the success of Mysis relicta introductions. Water 8098 Resources Research Institute, Laramie, Wyoming. Report 85, Boscarino BT, Rudstam LG, Tirabassi J, Janssen J, Loew ER (2010) 164 pp Light effects on alewife-mysid interactions in Lake Ontario: A Gregg RE (1976) Ecology of Mysis relicta in Twin Lakes, Colorado. combined sensory physiology, behavioral, and spatial approach. United State Department of Interior, Bureau of Reclamation, Limnology and Oceanography 55: 2061–2072, https://doi.org/ Denver, Colorado. Report REC-ERC-76-14, 70 pp 10.4319/lo.2010.55.5.2061 Britton LJ, Wentz DA (1980) Water-quality characteristics of Gregg RE, Bergersen EP (1980) Mysis relicta: effects of turbidity selected lakes and reservoirs in Colorado. United States and turbulence on short-term survival. Transactions of the Geological Survey (USGS), Lakewood, Colorado. USGS Open- American Fisheries Society 109: 207–212, https://doi.org/10.1577/ 1548-8659(1980)109<207:MR>2.0.CO;2 File Report 80-436, 145 pp Grossnickle NE (1982) Feeding habits of Mysis relicta – an overview. Brooks JL, Dodson SI (1965) Predation, body size, and composition Hydrobiologia 93: 101–107, https://doi.org/10.1007/BF00008103 of plankton. Science 150: 28–35, https://doi.org/10.1126/science.150. Hansen AG, Beauchamp DA (2015) Latitudinal and photic effects on 3692.28 Caldwell TJ, Wilhelm FM (2012) The life history characteristics, diel foraging and predation risk in freshwater pelagic growth and density of Mysis diluviana in Lake Pend O’Reille, ecosystems. Journal of Animal Ecology 84: 532–544, https://doi.org/10.1111/1365-2656.12295 Idaho, USA. Journal of Great Lakes Research 38: 58–67, Havel JE, Shurin JB (2004) Mechanisms, effects and scales of https://doi.org/10.1016/j.jglr.2011.07.010 Carim K, Wilcox T (2014) Protocol for collecting eDNA samples dispersal in freshwater zooplankton. Limnology and from streams- version 1.4. Rocky Mountain Research Station, Oceanography 49: 1229–1238, https://doi.org/10.4319/lo.2004.49.4_ part_2.1229 U.S.D.A. Forest Service, Missoula, Montana, 12 pp Havel JE, Lee CE, Vander Zanden JM (2005) Do reservoirs facilitate Carim KJ, Christianson KR, McKelvey KM, Pate WM, Silver DB, invasions into landscapes? BioScience 55: 518–525, Johnson BM, Galloway BT, Young MK, Schwartz MK (2016) https://doi.org/10.1641/0006-3568(2005)055[0518:DRFIIL]2.0.CO;2 Environmental DNA marker development with sparse biological Horne AJ, Goldman CR (1994) Limnology. 2nd edn. McGraw-Hill, information: A case study on Opossum Shrimp (Mysis diluviana). Inc., New York, 480 pp PLoS ONE 11: e0161664, https://doi.org/10.1371/journal.pone.0161664 Horppila J, Liljendahl-Numinen A, Malinen T, Salonen M, Dadswell MJ (1974) Distribution, ecology, and postglacial dispersal Tuomaala A, Uusitalo L, Vinni M (2003) Mysis relicta in a of certain crustaceans and fishes in eastern North America. eutrophic lake: consequences of obligatory habitat shifts. National Museum of Canada Publications in Zoology 11: 1–110 Limnology and Oceanography 48: 1214–1222, https://doi.org/10. Degraeve GM, Reynolds JB (1975) Feeding behavior and tempera- 4319/lo.2003.48.3.1214 ture and light tolerance of Mysis relicta in the laboratory. Idso SB, Gilbert RG (1974) On the universality of the Poole, Atkins Transaction of the American Fisheries Society 104: 394–397, Secchi disk-light extinction equation. Journal of Applied https://doi.org/10.1577/1548-8659(1975)104<394:FBATAL>2.0.CO;2 Ecology 11: 399–401, https://doi.org/10.2307/2402029 Devlin SP, Tappenbeck SK, Craft JA, Tappenbeck TH, Chess DW, Janiczek PMJ, DeYoung AJ (1987) Computer programs for sun and Whited DC, Ellis BK, Standord JA (2016) Spatial and temporal moon illuminance with contingent tables and diagrams. U.S. dynamics of invasive freshwater shrimp (Mysis diluviana): Naval Observatory Circular 171: 1–32, https://doi.org/10.21236/ Long-term effects on ecosystem properties in a large ADA182110 oligotrophic lake. Ecosystems 20: 183–197, https://doi.org/10.1007/ Johnson LE, Ricciardi A, Carlton JT (2001) Overland dispersal of s10021-016-0023-x aquatic invasive species: a risk assessment of transient Ellis BK, Stanford JA, Goodman D, Stafford CP, Gustafson DL, recreational boating. Ecological Applications 11: 1789–1799, Beauchamp DA, Chess DW, Craft JA, Deleray MA, Hansen BS https://doi.org/10.1890/1051-0761(2001)011[1789:ODOAIS]2.0.CO;2 (2011) Long-term effects of a trophic cascade in a large lake Jude DJ, Rudstam LG, Holda TJ, Watkins JM, Euclide PT, Balcer ecosystem. Proceedings of the National Academy of Sciences MD (2018) Trends in Mysis diluviana in the Great Lakes, 2006- 108: 1070–1075, https://doi.org/10.1073/pnas.1013006108 2016. Journal of Great Lakes Research (in press), https://doi.org/ Finnell LM (1972) Fryingpan-Arkansas fish research investigations. 10.1016/j.jglr.2018.04.006 Colorado Game, Fish, and Parks Division, Fort Collins. Progress Kestrup AM, Ricciardi A (2008) Occurrence of the Ponto-Caspian report, 24 pp mysid shrimp Hemimysis anomala (Crustacea, Mysida) in the Finnell LM (1977) Fryingpan-Arkansas fish research investigations. St. Lawrence River. Aquatic Invasions 3: 461–464, https://doi.org/ Colorado Division of Wildlife, Fort Collins. Progress report, 10.3391/ai.2008.3.4.17 104 pp Kolar CS, Lodge DM (2001) Progress in invasion biology: predicting Fürst M (1965) Experiments on the transplantation of Mysis relicta invaders. Trends in Ecology and Evolution 16: 199–204, Lovén into Swedish lakes. Institute of Freshwater Research https://doi.org/10.1016/S0169-5347(01)02101-2 Drottningholm Report 46: 79–87 Lasenby DC, Northcote TG, Fürst M (1986) Theory, practice, and Fredrickson L (2017) Bio-invasions and bio-fixes: Mysis shrimp effects of Mysis relicta introductions to North American and introductions in the twentieth century. Environment and History Scandinavian lakes. Canadian Journal of Fisheries and Aquatic 23: 285–320, https://doi.org/10.3197/096734017X14900292921789 Sciences 43:1277–1284, https://doi.org/10.1139/f86-158

419 B.M. Johnson et al.

Laske SM, Rahel FJ, Hubert WA, Cavalli PA (2011) Ecology of Schoen ER, Beauchamp DA, Buettner AR, Overman NC (2015) unique lentic populations of Roundtail Chub, Gila robusta. Temperature and depth mediate resource competition and Western North American Naturalist 71: 507–515, https://doi.org/ apparent competition between Mysis diluviana and Kokanee. 10.3398/064.071.0408 Ecological Applications 25: 1962–1975, https://doi.org/10.1890/14- Leopold LB (1980) Bathymetry and temperature of some glacial 1822.1 lakes in Wyoming. Proceedings of the National Academy of Scott WB, Crossman EJ (1973) Freshwater fishes of Canada. Sciences 77: 1754–1758, https://doi.org/10.1073/pnas.77.4.1754 Fisheries Research Board of Canada Bulletin 184, Ottawa, Martinez PJ (2012) Invasive crayfish in a high desert river: Ontario, Canada, 966 pp Implications of concurrent invaders and climate change. Aquatic Sell DW, Evans MS (1982) A statistical analysis of subsampling and Invasions 7: 219–234, https://doi.org/10.3391/ai.2012.7.2.008 an evaluation of the Folsom plankton splitter. Hydrobiologia 94: Martinez PJ, Bergersen EP (1989) Proposed biological management 223–230, https://doi.org/10.1007/BF00016403 of Mysis relicta in Colorado lakes and reservoirs. North Sherman RK, Lasenby DC, Hollet L (1987) Influence of oxygen American Journal of Fisheries Management 9: 1–11, concentration on the distribution of Mysis relicta Loven in a https://doi.org/10.1577/1548-8675(1989)009<0001:PBMOMR>2.3.CO;2 eutrophic temperate lake. Canadian Journal of Zoology 65: Martinez PJ, Bergersen EP (1991) Interactions of zooplankton, Mysis 2646–2650, https://doi.org/10.1139/z87-401 relicta, and kokanees in Lake Granby, Colorado. American Silver DB, Johnson BM, Pate WM, Christianson KR, Tipton J, Fisheries Society Symposium 9: 49–64 Sherwood J, Smith B, Hao Y (2016) Effect of net size on Martinez PJ, Bigelow PE, Deleray MA, Fredenburg WA, Hansen estimates of abundance, size, age, and sex ratio of Mysis BS, Horner NJ, Lehr SK, Schneidervin RW, Tolentino SA, diluviana. Journal of Great Lakes Research 42: 731–737, Viola AE (2009) Western Lake Trout woes. Fisheries 34: 424– https://doi.org/10.1016/j.jglr.2016.02.012 442, https://doi.org/10.1577/1548-8446-34.9.424 Simberloff D, Von Holle B (1999) Positive interactions of non- Martinez PJ, Gross MD, Vigil EM (2010) A compendium of indigenous species: invasional meltdown? Biological Invasions crustacean zooplankton and Mysis diluviana collections from 1: 21–32, https://doi.org/10.1023/A:1010086329619 selected Colorado reservoirs and lakes, 1991-2009. Fort Collins, Spencer CN, McClelland BR, Stanford JA (1991) Shrimp stocking, CO: Colorado Division of Wildlife, 278 pp salmon collapse and eagle displacement: cascading interactions Minchin D, Boelens R (2010) Hemimysis anomala is established in in the food web of a large aquatic ecosystem. Bioscience 41: 14– the Shannon River Basin District in Ireland. Aquatic Invasions 5: 21, https://doi.org/10.2307/1311536 S71–S78, https://doi.org/10.3391/ai.2010.5.S1.016 Stohlgren TJ, Jarnevich CS (2009) Risk assessment of invasive Morgan MD, Threlkeld ST (1982) Size dependent horizontal migra- species. In: Clout MN, Williams PA (eds), Invasive species tion of Mysis relicta. Hydrobiologia 93: 63–68, https://doi.org/ management. Oxford University Press, New York, pp 19–35 10.1007/BF00008099 Van Duyn-Henderson JA, Lasenby DC (1996) Zinc and cadmium Nesler T (1986) Mysis-gamefish studies. Job Progress Report, project transport by the vertically migrating Opossum Shrimp, Mysis F-83-R. Fort Collins, CO: Colorado Division of Wildlife, 99 pp relicta. Canadian Journal of Fisheries and Aquatic Sciences 43: Nesler TP, Bergersen EP (1991) Mysids in fisheries: hard lessons 1726–1732, https://doi.org/10.1139/f86-216 from headlong introductions. American Fisheries Society Vander Zanden MJ, Olden JD (2008) A management framework for Symposium 9: 1–4 preventing the secondary spread of aquatic invasive species. Pejchar L, Mooney HA (2009) Invasive species, ecosystem services Canadian Journal of Fisheries and Aquatic Sciences 65: 1512– and human well-being. Trends in Ecology and Evolution 24: 1522, https://doi.org/10.1139/F08-099 497–504, https://doi.org/10.1016/j.tree.2009.03.016 WGFD (2017) Wyoming Game and Fish Department (WGFD) Peterson DA, Averett RC, Mora KL (1987) Water quality of Wyoming fishing regulations. https://wgfd.wyo.gov/Regulations/Regu Fremont Lake and New Fork Lakes, western Wyoming- a lation-PDFs/WYFISHINGREGS_BROCHURE (accessed September progress report. U.S. Geological Survey, Denver, Colorado. 2017) Water-Resources Investigations Report 86-4016, 55 pp WDEQ (2015) Wyoming Department of Environmental Quality Pimental D, Zuniga R, Morrison D (2005) Update on the environ- (WDEQ) Water quality condition for Brooks Lake, mental and economic costs associated with alien-invasive Wind/Bighorn Basin, 2009, 2011-2012. Water Quality Division, species in the United States. Ecological Economics 52: 273–288, Cheyenne, 101 pp https://doi.org/10.1016/j.ecolecon.2004.10.002 Wittmann KJ, Ariani AP (2009) Reappraisal and range extension of Ricker KE (1959) The origin of two glacial relict crustaceans in non-indigenous Mysidae (Crustacea, Mysida) in continental and North America, as related to Pleistocene glaciation. Canadian coastal waters of eastern France. Biological Invasions 11: 401– Journal of Zoology 37: 871–893, https://doi.org/10.1139/z59-085 407, https://doi.org/10.1007/s10530-008-9257-7 Sandeman IM, Lasenby DC (1980) The relationships between ambient oxygen concentration, temperature, body weight, and oxygen consumption for Mysis relicta (Malacostraca: Mysidacea). Canadian Journal of Zoology 58: 1032–1036, https://doi.org/10. 1139/z80-145

420